Trace the Original Carbon in Soil Carbonate and to Correct the "Dead" Car- Bon Dilution Effect Is an Important Concern of 14C Dating

Home , Nodule (geology)

[RADIOCARBON, VOL 28, No. 2A, 1986, P 464-472] VALIDITY OF 14C AGES OF CARBONATES IN SEDIMENTS YIJIAN CHEN 14C Dating Laboratory, Institute of Geology, State Seismological Bureau, Beijing, China and HENRY POLACH Radiocarbon Dating Research, Australian National University, Canberra

ABSTRACT. This review is based on geologic surveys carried out in Australia and China as well as on more than 300 14C dates published in Radiocarbon. Evaluated are the origins and pathways of carbonate formation, stable isotopic composition, carbonate nodule growth rates and paleo-climatic effects. The three identified 613C abundance peaks are unrelated to envi- ronment and carbon source whilst 14C ages group themselves into periods corresponding to past humid warm climate. It is concluded that the major error in caliche dating is due to incorporation of old limestone whilst error on nodule dating is related to their slow growth rate. Thus, caliche antedates and nodules postdate the times of their deposition. Delta 13C values cannot be used to correct for limestone or atmospheric contamination effects.

INTRODUCTION There are various types of carbonate precipitates (authigenic carbonate) in soils, lake deposits, marine sediments, and loess that can be used for 14C dating when there is a lack of organic carbon. However, the validity of authigenic carbonate dates are suspect since the initial 14C concentration is unknown and post-depositional contamination/alteration is often possible. 14C determinations on this type of material - both in Australia and China - have mostly been interpreted successfully where stratigraphic relation- ships have been set up with dated sequences and when associated environmental factors have been recognized. This paper is based on the assessment of the results of geologic surveys carried out in Australia and China and incorporates more than 300 14C dates, most of which have been published in Radiocarbon. It discusses the problems associated with the validity of soil carbonate as a dating material.

CARBON ISOTOPIC COMPOSITION Soil carbonate formation is generally considered to result from carbonate supersaturation of soil solution due to evaporation, evapotranspira- tion, and lowered p(C02). There are four possible carbon sources of soil authigenic carbonate: 1) from soil C02, 2) from dissolved carbonate in ground water, 3) from old calcareous dust derived from eroded soil expo- sure, 4) from limestone or shell detritus in parent sediments (Geyh, 1970; Leamy & Rafter, 1972; Pantin, 1963; Salomons & Mook, 1976). Stable isotopes in soil carbonates are useful tracers for understanding the process and circumstance of soil carbonate formation (Cerling, 1984; Salomons, Goudie & Mook, 1978). Whether b13C values can be used to trace the original carbon in soil carbonate and to correct the "dead" carbon dilution effect is an important concern of 14C dating. b13C values of soil carbonate show values (Fig 1) ranging from 0%o to -13%o and falling into three sections, 0%o to - 2%o, - 3%o to - 6%o, and - 7%o to -11 %o. 464 Validity of '4C Ages of Carbonates in Sediments 465

40 i 35 U

30 " 25 0 20 L

E z 15

2 -4 -6 - 3 -10 12 14

Pedogenic carbonate ) 13c (` I Fig I. 1)isitribution of 'C values of caliche and calcareous nodules

b13C values of soil CO2 in temperate and semi-arid climate also shows a wide range (Fig 2). In general, the carbon isotopic compositions of soil CO2 is related to the proportion of biomass using the C4 photosynthetic pathway and the CO2 respiration rate of the soil; at low soil respiration rates signifi- cant atmospheric C02 mixing can occur. b13( values of groundwater HCO3 also show a wide range. 8130 values of rain are generally considered to range from -14%o to - I8%o In the theoretical calculation, fractionation factors (E) given by Hendy (1971) are adoptedl for calculating the anticipated b13C values. If soil carbonate was formed by weathering of feldspar (hook, 1976), 8130 assuming values of soil CO2 range from -12%o to - 25%o. Conse- 8130 quently, the anticipated values range from - 2.5%o to -15.5%o (see Fig 2). If soil carbonate was formed from the dissolution of limestone detritus b13C (b13C values of 0%o) by rain, the anticipated values will range from - 7.9%o to -10.9%o. 466 Yijian Chen and Henry Polach

SOIL ATMOSPHERE CO 2 Atmosphere Arizona ,USA Florida , USA Texas ,USA Ontario Canada GROUNDWATER Salomons et al , (1978) Saudi Arabia London ,UK Lancashire, UK Kufra,L ibya Triassic Aguifer

Brazil PEDOGENIC CARBONATE Long Island, USA Holland Spain Italy Australia India Libya Austria Texas USA Cyprus Kenya F rance South Africa China Nigeria Netherlands PEDOGENIC CARBONATE (VALUES OF THEORETICAL CALCULATIONS

Soil CO2 Groundwater Rain + Limestone

-2 -4 -6 -8 -10 -12 -14 -16 -18 -20 -22 -24 -26 -28 -30 13C Iuol Fig 2. Distribution of 13C values of soil atmosphere CO2, groundwater HCO3, and authigenic carbonate in caliche and calcareous nodules

If soil carbonate was formed by reprecipitation of dissolved carbonate in ground water, the anticipated 5'3C values will range from -3.9%o to 12.9%o. Because anticipated 5'3C values of different original soil carbonate are analogous, it is impossible to trace carbon sources of soil carbonate from 5'3C values. In addition, S13C values of soil carbonate change while redissolution and reprecipitation occur. Hence, 3'3C values are not suitable for correct- ing the "dead" carbon dilution effect on soil carbonate dates. Bowler and Polach (1971) have indicated that 13C/12C ratios remain relatively constant despite variations in soil types, carbon sources, vegeta- tive cover, climate, and carbonate mineralogy. This demonstrates that using S13C values to trace carbon sources is rather difficult. Validity of 14C Ages of Carbonates in Sediments 467

AGE COMPARISON WITH ORGANIC CARBON To test the reliability of authigenic carbonate dates as a time scale, the authigenic carbonate dates are compared with dates determined for those of co-existing organic carbon. Figure 3 shows 82 pairs of dates plotted from determinations carried out by different laboratories around the world. Three important conclusions may be drawn: 1) approximately half of the carbonate dates are older and half are younger than the organic carbon dates, 2) ca 90% of the pairs differ from each other by 1000 to 3000 years and 10% differ from each other by 3000 to 20,000 yr, 3) the majority of dates from drill cores, obtained from lake deposits or marine sediments, differ by <1000 yr. Williams and Polach (1971) have proposed that the segregation of authigenic carbonate occurs largely after the accumulation of the coexist- ing organic carbon in the host material; thus, the carbonate dates should be younger than the organic carbon dates. Accordingly, if dates for carbonates are older than those for organics this will suggest that incorporation of carbonate, containing little or no 14C, eg, a "limestone dilution" has occurred

T 5 10 15 20 25 30

14C age (Organic fraction) x 103years Fig 3. Age comparison of sample pairs of carbonate in sediments and co-existing organic carbon 468 Yijian Chen and Henry Polach during formation. The contribution of "dead" carbon to soil carbonate is estimated to be 10%-30% of the total carbon, based on the observed "limestone dilution effect" shown in Figure 3. This is in basic agreement with Goodfriend and Hood (1983) who estimated that the contribution of limestone carbon to snail shell is 0% to 33%.

GROWTH RATE AND APPARENT AGE The formation of soil carbonate involves a three-stage process: 1) provision of carbonate-rich solution, 2) the movement of that solution, 3) the precipitation of soil carbonate (Salomons, Goudie & Mook, 1978). Because the provision and movement of carbonate-rich solution in soil are slower than the tufa deposited by a spring or travertine formed in a cave, we assumed that the growth rate of soil carbonate in arid and semi- arid zones is rather slow. Using 14C techniques, Peterson (1966) reported that dolomite crystals in playa deposits grow at the rate of 0.46-0.86 µ/Myr. By measuring caliche profiles in Texas (Fig 4), Valastro, I)avies and Rightmire (1968) suggested that the dates closest to the surface are 16,000 to 25,000 yr, changing with increasing depth (to 60cm) to 30,000-35,000 yr. Dating the outer and inner portions of calcareous nodules revealed (Rafter et al, 1972) age/time differences of 280-16,400 yr (Table 1). Callen, Wasson and Gillespie (1983) reported that the carbonate segregation in Moko Palaeosol, Australia took at least 4000-15,000 yr to form. This evidence demonstrated that the apparent age of soil carbonate only represents the greatest minimum age. The soil carbonate age is considered younger than the age of parent sediment accumulation, so that the slower the growth rate of the soil carbonate, the greater the deviation from the accumulation time of the parent sediment.

POST-DEPOSITIONAL ALTERATIONS There are several processes that cause post-depositional alterations: recrystallization, replacement, overgrowth, and dolomitization. Thin- section studies of calcareous nodules (Williams & Polach, 1971) indicated that authigenic carbonate is commonly comprised of microcrystalline calcite, ca l0µ across, which encloses detrital grains of quartz and feldspar. However, in many nodules the carbonate groundmass is significantly coarser grained around the quartz and feldspar as well as in the voids, with individual crystallites attaining 20,i across. The nodules composed of coarser crystallite give anomalously younger ages, showing that recrystallization and contamination with 14C have occurred in the late formation stage. Chemical and x-ray diffraction analyses of concretions formed in the recent supratidal flats of Broad Sound, Queensland (Cook & Polach, 1973) indicated that concretions up to 20cm are composed of micritic low-mag- nesium calcite and dolomite. However, poorly crystalline dolomite concen- trates mainly in the outer portion of the concretions suggesting that it was formed by post-depositional dolomitization of the calcitic host material. Bowler and Polach (1971) observed that the percentage of dolomite increases progressively with age in paleosols at Nyah, western Australia. All TABLE 1 Comparison 14 ages of calcareous nodules between inner and outer layers Growth time t4C Species Sampling site age ( Yr BP) (Yr) Ref SYngenetic Atlantic, water depth inner: 690 ± 50 400 ± 60 Vogel (1970) nodule 73m, 6cm diameter (1) Outer: 290 ± 40 Inner: 1000 ± 45 2 80 ± 5 5 (2) Outer: 720 ± 30 Atlantic, 91 m Inner: 26,640 ± 600 Hubby, Biers and Suess (1965) Middle: 19,440 ± 600 8980 ± 750 Whole: 17,660 ± 450 Atlantic, water depth Inner: 1050 ± 100 1050 ± 100 Delibrias, Guiltier and LabeYrie 89m, 5cm diameter Outer: Modern (1972) East China Sea, water depth Inner: 9625 ± 130 2845 ± 170 14C dates Inst Geol Natl Bur 163m, 6cm in diameter Outer: 6780 ± 110 Seismology (ms) Pcdogcnic New Zealand nodule Large nodule (1) Inner: 9900 1690 Rafter et al (1972 Outer: 8210 Small nodule (2) Inner: 7070 670 Outer: 6400 Circular nodule (3) Inner: 34,400 ± 900 16,400 ± 950 Outer: 18,000 ± 300 Illinois, USA Inner: 7370 ± 140 Coleman (1974) nodule Middle: 5865 ± 80 2340 ± 170 Outer: 5030 ± 100 Australia Rim: 29,000 Callers, Wasson and Gillespie nodule Center: 44,000 15,000 (1983) Whole: 39,800 ± 3000 Shandong, China Inner: 29,530 ± 790 9070 ± 850 14C dates Inst Geol Natt Bur nodule Outer: 20,460 ± 300 Seismology (ms) Najing, China Inner: 40,000 ± ? 9400 ± 110 Li and Fang (1983) p nodule Outer: 30,600 ± 1100 470 Yijian Chen and Henry Polach

14 C age (x103 years)

30 35 40 115 20 25

Fig 4. 14C date distribution in caliche from western Texas, (after Valastro, Davies & Rightmire, 1968) dolomite samples in this area showed considerably higher levels of 14C activ- ity than their true ages would allow, suggesting dolomitization associated with cation/anion exchanges with active carbon. Both the rarity of rain in arid regions and the impermeability of the dense and hard calcareous nodules seem to be beneficial factors for retard- ing post-depositional alterations (Williams & Polach, 1969). This is shown by 1) the apparent lack of dates for and zone soil carbonate youner than 1000 yr (see Fig 5), 2) the existence of soil carbonate with very low 4C con- tent, and 3) the dates of soil carbonate in arid regions seem to relate to the late stages of soil formation. These observations are based on the assump- tion that the uptake of young carbon occurs more readily under wet condi- tions.

CLIMATIC INFLUENCE Soil and paleosols with accumulations of authigenic carbonates are restricted to arid and semi-arid regions. The detailed mechanisms of carbonate formation in soil and loess are not well known, but it is believed that climatic factors play a major role. Frequency distributions of 14C dates based on 327 determinations on caliches and calcareous nodules in different parts of the world have been plotted in Figure 5. Seven peaks, corresponding to well-established climatic 14C Validity of Ages of Carbonates in Sediments 471

0 10 15 20 25 30 35 >35 14C age (x103 years) °a c°a a' am a, j yam Q ' , .cJ a) a

a me

J} yV°,° c .J \ Vm45 ci Fig 5. Frequency distributions of 14C ages of caliche (shaded) and calcareous nodules (light area) from different parts of the world

(humid-warm) periods (Cao, 1983), can be identified with the exception of the fifth period showing the Wurm and Wisconsin loess deposition period in Europe and North America which are believed to be cold and dry. Stud- ies on paleosols developed in loess reveal that some intermittent humid- warm stage has occurred during this period (Liu, 1964; Ruhe, Miller & Vreeken, 1970). The hypothesis that caliche and calcareous nodules are developed during relatively humid-warm periods seems to be confirmed by the available 14C dates.

CONCLUSIONS Radiocarbon dates of carbonate from lacustrine or marine sediments generally agree with co-existing organic material, but soil carbonate dates significantly deviate from the co-existing organic carbon dates. In some cases, soil carbonate is precipitated from a solution in which 14C concentration is not in equilibrium with soil CO2, so that some "dead" carbon from ground water or older carbon may be incorporated. Age comparison with organic material estimates the contribution of "dead" carbon to soil carbonate at 10%-30% of the total carbon. In general, the segregation of soil carbonate takes thousands of years. Thus, the 14C age of soil carbonate is younger than its parent sediments and can be regarded as the greatest minimum age of segregation. Conse- quently, soil carbonate dating does not provide reliable estimates of the age of pedogenesis, but can be useful if its limitations are recognized. 472 Yijian Chen and Henry Polach ACKNOWLEDGMENTS This paper was written while one of the authors (Chen) was a Visiting Fellow at the Australian National University Radiocarbon Dating Research Laboratory. The manuscript was read and valuable criticism offered by Aro Arakel, Queensland Institute of Technology, Brisbane. All staff of the Radiocarbon Laboratory, ANU, are thanked for their support and help during Yijian Chen's visit.

REFERENCES

Bowler, J M and Polach, H A, 1971, Radiocarbon analyses of soil carbonates; an evaluation from paleosols in southeastern Australia, in Yaalon, D H, ed, Paleopedology-origin, nature and dating of paleosols: Jerusalem, Israel Univ Press, p 97-108. Cao, J, 1983, Quarternary geology: Beijing, Shang Wa Press, (in Chinese). Callen, R A, Wasson, R J and Gillespie, R, 1983, Reliability of radiocarbon dating of pedo- genie carbonate in the Australian arid zone: Sed Geol, v 35, p 1-14. Cerling, T E, 1984, The stable isotopic composition of modern soil carbonate and its relation- ship to climate: Earth Planetary Sci Letters, v 71, p 229-240. Coleman, D I), 1974, Illinois State Geological Survey radiocarbon dates V: Radiocarbon, v 16, p 105-117. Cook, P J and Polach, H A, 1973, Discovery of recent supratidal dolomite at Broad Sound, Queensland: Search, v 4, no. 3, p 78-79. Delibrias, G, Guillier, M T and Labeyrie, J, 1972, Gif natural radiocarbon measurements VII: Radiocarbon, v 14, p 280-320. Geyh, M A, 1970, Carbon-14 concentration of lime and soil and aspects of the carbon-14 dating of groundwater, in Isotope hydrology: Vienna, IAEA, p 213-225. Goodfriend, G A and Hood, D G, 1983, Carbon isotope analysis of land snail shells: Implica- tions for carbon sources and radiocarbon dating: Radiocarbon, v 25, p 810-830. Hendy, C H, 1971, The use of 14C in the study of cave processes, in Olsson, I U, ed, Radiocar- bon variations and absolute chronology, Nobel symposium, 12th, Proc: Stockholm, Almqvist & Wiksell, p 419-443. Hubbs, C I,, Bien, G S and Suess, H E, 1965, La Jolla natural radiocarbon measurements IV: Radiocarbon, v 7, p 66-117. Leamy, M A and Rafter, T A, 1972, Isotope ratios preserved in pedogenic carbonate and their application in paleopedology, in Rafter, T A and Grant-Taylor, T, eds, Internatl 14C conf, 8th, Proc: Wellington, Royal Soc New Zealand, p D42-D48. Li, L and Fang, Y, 1983, Studies on the origin and epoch of calcareous nodules within loess: Nanjing, Quaternary geol peat research, Nanjing Educ Col! (in Chinese), p 10-14. Liu, D S, 1964, The loess in the middle of the Yellow River: Beijing, Science Press, (in Chi- nese). Monk, W G, 1976, The dissolution-exchange model for dating groundwater with 14C, in Inter- pretation of environmental isotope and hydrochemical data in groundwater hydrology: Vienna, IAEA, p 213-225. Pantin, H M,1963, Rate of formation of a diagenetic calcareous concretion: Jour Sed Petrol, v 28, no. 3, p 366-371. Peterson, M N, 1966, Growth of dolomite crystals: Am Jour Sci, v 264, p 257-272. Rafter, A T, Jansen, H S, Lockerbie, L and Trotter, M M, 1972, New Zealand radiocarbon reference standards, in Rafter, T A and Grant-Taylor, T, eds, Internatl 14C conf, 8th, Proc: Wellington, Royal Soc New Zealand, p H29-H80. Ruhe, R V, Miller, G A and Vreeken, W G, 1970, Paleosols, loess sedimentation and soil stra- tigraphy, in Yaalon, D H, ed, Paleopedology origin, nature and dating of paleosols: Jeru- salem, Israel Univ Press, p 41-60. Salomons, W and Mook, W G, 1976, Isotope geochemistry of carbon dissolution and reprecipitation in soil: Soil Sci, v 122, p 15-24. Salomons, W, Goudie, A and Mook, W G, 1978, Isotopic composition of calcrete deposits from Europe, Africa and India: Earth Surface Processes, v 3, p 43-57. Valastro, 5,1)avies, E M and Rightmire, CT, 1968, University of Texas at Austin radiocarbon dates VI: Radiocarbon, v 10, p 384-401. Vogel, J C, 1970, Groningen radiocarbon dates IX: Radiocarbon, v 12, p 444-471. Williams, G E and Polach, H A, 1969, The evaluation of 14C ages for soil carbonate from the arid zone: Earth Planetary Sci Letters, no. 7, p 240-242. 1971, Radiocarbon dating of arid-zone calcareous paleosols: Geol Soc America Bull, v 82, p 3069-3086.